Geochemcial Tracing of Compartmentalization and Efficiency of Water Injection in the Junggar Basin

Date of Award


Degree Type


Degree Name

Master of Science (MS)


Earth Sciences


Donald I. Siegel


China, Fluorescence, Injection and Production wells, Mixing Models, Multivariate Statistics, Tracers

Subject Categories

Earth Sciences


The paper presents the results of using natural tracers in the Junggar Basin, western China to characterize the effectiveness of injected water in displacing oil and to find where injected water moves in high permeability zones. I collected 91 samples for solute analysis and 94 samples for fluorescence in three exploration areas, producing oil from Jurassic and Triassic age fluvial and deltaic sediments in three blocks; Block 29 (Triassic T2k2, T2k1 and Jurassic J1b formation); Block 60 (Jurassic J3q formation), and Block 67 (Jurassic J1b formation) in the Hongshanzui and Chepaizi oilfield in the western part of the basin).

From mixing models using bromide concentrations in injected and produced water, injected water in Block 29 T2k2 formation defines a long plume, which could reflect transport along depositional narrow sand bar-like deposits within the fluvial sediment. In the T2k1 formation in Block 29, three flow paths appear to occur, two trending south to north and one from east to west. In Block 29, J1b formation, and Block 67, J1b formation, water spreads in all directions from injection wells. In block 60, production from wells from multiple formations made using natural tracers ineffective.

Several key natural background fluorescence peaks produced from dissolved organic substances (e.g. ë286 and ë618 nm) varied linearly in block 29, Triassic T2k2 formation. Block 29, Triassic T2k1 formation appeared to have, based on fluorescence, two possible kinds of formation waters, one defined by fluorescence peaks ë290 and ë324 and another also by peaks ë571 and ë621. Block 60 produces water from multiple formations, and fluorescence data could not clearly define end members because mixing included waters from multiple formations.

Two component-mixing models based on major fluorescence peaks defined contour maps different from those produced from bromide. Perhaps BTEX in the formation waters degraded to produce relative fluorescence intensity that reflects degradation and organic acids, rather than organics in the formation water itself. Or perhaps the injected water solubilizes organics from the oil in the formation, thereby masking diagnostic fluorescence peaks. Multiple fluorescence peaks also overlap and complicate the analysis from a quantitative standpoint.

Correlation Matrix statistical analysis, Principal Component Analysis (PCA), Discriminant Analysis (DA) and Cluster Analysis (CA) can, however distinguish among wells from background fluorescence peaks. Well H0538 and H0214 in block 29 Triassic T2k2, well H0218, H0225 and H0208 in block 29 Triassic T2k1, well H0510 and H0702 in block 29 Triassic J1b, Well H8031 and H8021 in block 67 Jurassic J1b have different dominant fluorescent peak combinations than other wells in those blocks. Discriminate analysis shows fluorescence of water from the J3q formation differs from the J1b, T2k1 and T2k2 formations.

Statistical analysis of solute concentrations does not show the same results as BFA peaks did. The major solutes mainly show which major elements dominate in the area and wells. Much sulfate exists in the injected water, which can combine with Ca and may block the pores in the formation by precipitation of calcium sulfate.

Characterizing waters from different formations would be useful to determine the extent to which formations actually are isolated or not from each other. At least using natural tracers, such as bromide, avoids complications of introducing radioactive substances or other anthropogenic tracers into otherwise natural systems, although it appears that BFA is not a reliable tracer method for this application.

This document is currently not available here.